From ceb3779bbe150793494820ec42d44b5f771778ee Mon Sep 17 00:00:00 2001
From: Davies Liu <davies@databricks.com>
Date: Sat, 18 Apr 2015 00:36:53 -0700
Subject: [PATCH] Update rdd.py

---
 python/pyspark/rdd.py | 1 +
 1 file changed, 1 insertion(+)

diff --git a/python/pyspark/rdd.py b/python/pyspark/rdd.py
index 73f7cbac2f6a0..d254deb527d10 100644
--- a/python/pyspark/rdd.py
+++ b/python/pyspark/rdd.py
@@ -2268,6 +2268,7 @@ def _prepare_for_python_RDD(sc, command, obj=None):
         broadcast = sc.broadcast(pickled_command)
         pickled_command = ser.dumps(broadcast)
     # There is a bug in py4j.java_gateway.JavaClass with auto_convert
+    # https://github.com/bartdag/py4j/issues/161
     # TODO: use auto_convert once py4j fix the bug
     broadcast_vars = ListConverter().convert(
         [x._jbroadcast for x in sc._pickled_broadcast_vars],